[GitHub] [pinot] Jackie-Jiang commented on pull request #11380: Properly handle shutdown of TableDataManager

[via GitHub]

Jackie-Jiang commented on PR #11380:
URL: https://github.com/apache/pinot/pull/11380#issuecomment-1684240370

   > This looks good! One scenario we might need to check for is creating a 
table just after it is deleted. We ran into this problem before.
   > 
   > <img alt="Screenshot 2023-08-18 at 2 14 16 PM" width="917" 
src="https://user-images.githubusercontent.com/5559703/261552776-50d58a97-3641-4871-84cd-f5b3bbaf5eb4.png";>
   
   Correct. Currently we can still run into this problem, but it will be easier 
to catch because we don't allow adding segment to the already shut down table 
data manager. To solve this problem, we should check if server has successfully 
dropped the table. Let me see if it can be easily added. If not, we can address 
it separately.
   
   As for the `stop` or `preShutDown`, I decided to not add them because the 
contract around them is unclear to me. Should we block other operations after 
`stop` is called? Or only `stop` the upsert metadata manager but allow all 
other operations? If we block other operations, then there is no difference 
between `stop` and `shutDown`; if we do not block other operations, I don't 
know the side effect of adding another segment after `stop`, and it can run 
into very weird state if somehow `stop` is called but `shutDown` is not (some 
exception happened in between)


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org
For additional commands, e-mail: commits-h...@pinot.apache.org